Full Terms & Conditions of access and use can be found athttp://www.tandfonline.com/action/journalInformation?journalCode=upcp20

Political Communication

ISSN: 1058-4609 (Print) 1091-7675 (Online) Journal homepage: http://www.tandfonline.com/loi/upcp20

Changing Owners, Changing Content: Does WhoOwns the News Matter for the News?

Allison M. Archer & Joshua Clinton

To cite this article: Allison M. Archer & Joshua Clinton (2017): Changing Owners, ChangingContent: Does Who Owns the News Matter for the News?, Political Communication, DOI:10.1080/10584609.2017.1375581

To link to this article: https://doi.org/10.1080/10584609.2017.1375581

View supplementary material

Published online: 10 Nov 2017.

Submit your article to this journal

Article views: 150

View related articles

View Crossmark data

Changing Owners, Changing Content: Does WhoOwns the News Matter for the News?

ALLISON M. ARCHER and JOSHUA CLINTON

The press is essential for creating an informed citizenry, but its existence depends onattracting andmaintaining an audience. It is unclear whether supply-side effects—includingthose dictated by the owners of the media—influence how the media cover politics, yet thisquestion is essential given their abilities to set the agenda and frame issues that are covered.We examine how ownership influences media behavior by investigating the impact ofRupertMurdoch’s purchase of the Wall Street Journal (WSJ) in August 2007. We collectdata on every front-page story and editorial for 27 months, and we compare thedifference in political coverage between the New York Times (NYT) and WSJ using adifference-in-differences design. We show that the amount of political content in theopinion pages of the two papers were unchanged by the sale, but the WSJ’s front-pagecoverage of politics increased markedly relative to the NYT. Similar patterns emergewhen comparing the WSJ’s content to USA Today and the Washington Post. Our findinghighlights potential limits to journalists’ ability to fulfill their supposed watchdog rolein democracies without interference from owners in the boardroom.

Keywords Media owners, newspapers, political news

The media occupy a curious position in society. On the one hand, a free press is thought tobe essential for a democratic and well-functioning society, as it is through the press that theactions of the government and political elites can be revealed and held accountable. On theother hand, the media are profit-maximizing businesses that aspire to maintain and growtheir audiences and that often report to owners concerned with profitability. No longersubsidized by political parties—and with a competitive environment that allows consu-mers to select nearly any media outlet they desire—media companies must create brandsfor which consumers are willing to pay (Hamilton, 2004).

The extent to which there is tension between the aspirational and economicmotivations ofthe press is unclear. We explore this tension by investigating the impact of ownership oncoverage decisions; when a new owner takes over a paper, does its political coverage changeas a consequence? It may seem obvious that ownership matters for institutions whose coveragedecisions define the brand being sold to consumers, but systematic characterizations of theimpact of changing ownership are based largely on anecdotes or personal experiences.

Allison M.N. Archer is an assistant professor in the Jepson School of Leadership Studies,University of Richmond. Joshua D. Clinton is the Abby and Jon Winkelried Chair and professor ofPolitical Science at Vanderbilt University.

Address correspondence to Joshua Clinton, PMB 505, 230 Appleton Place, Nashville, TN37203-5721. E-mail: josh.clinton@vanderbilt.edu

Color versions of one or more of the figures in the article can be found online at www.tandfonline.com/UPCP.

Political Communication, 00:1–18, 2017Copyright © 2017 Taylor & Francis Group, LLCISSN: 1058-4609 print / 1091-7675 onlineDOI: https://doi.org/10.1080/10584609.2017.1375581

1

Investigating the impact of ownership on coverage decisions is important because it iswell-known that media can set both the social and political agenda (Iyengar & Kinder,1987). In fact, media firms represent valuable entities with high “amenity potential” giventhe possibility that owners can systematically influence public opinion through their newsoutlets (Demsetz & Lehn, 1985). Understanding whether political coverage depends on whoowns the paper is critically important for assessing the media’s ability to inform citizensabout the actions of their government, but few studies examine the factors that determinewhich content makes its way to the top of reporters’ and editors’ agendas (Nelson, Bryner, &Carnahan, 2011; but see, e.g., Dunaway, 2008; Larcinese, Snyder, & Puglisi, 2011; PewResearch Center, 2011; Wagner & Collins, 2014). To contribute to our understanding, wecollect data on the content and emphasis given to stories by leading newspapers tocharacterize how the coverage of politics changes in response to a change in ownership.

In particular, we examine the impact of the sale of the Wall Street Journal (WSJ) toRupert Murdoch on August 1, 2007, by collecting data on the daily content and coverageof the WSJ and the New York Times (NYT) over 27 months. We compare changes in thefront-page and editorial content of the WSJ to that of the NYT over time to identify whetherchanging patterns in political coverage by the WSJ can be attributed to Murdoch ratherthan changes in the news cycle or media environment. We further compare the increasedpolitical coverage in the WSJ after the transfer of ownership we document to the coveragegiven by the USA Today (USAT) and the Washington Post (WaPo), two papers that differin the amount of attention they traditionally give to politics.

Our primary interest is whether who controls the media affects the amount of attentiongiven to political issues, and we find clear evidence that it does. The editorial pages ofboth the WSJ and NYT were equally likely to address political issues throughout theexamined time period, but the WSJ became increasingly more likely to feature politicalstories on its front pages following Murdoch’s purchase than the NYT, USAT, and theWaPo. Measuring political coverage using the percentage of political stories on the frontpage and the percentage of political stories above the fold reveals that the front-pagecoverage of the WSJ—which had historically focused less on politics than the NYT—wasonly slightly less focused on politics than the NYT following its sale.1

The fact that the amount of political coverage changed drastically only for the WSJ andthat the increase only occurred after the sale of the WSJ strongly suggests that who owns themedia matters. While some of the larger implications of this effect cannot be discerned fromthis analysis—for example, what the motivations were for the change we identify; howexactly the coverage changed; the impact of the changes on the opinions of existing and newsubscribers—the results unambiguously demonstrate that the ownership of news organiza-tions matters for the news that gets covered. Better understanding the incentives of thoseresponsible for managing and producing the news is, therefore, important for assessing therole of the media in a democracy; if media coverage depends at least in part on who ownsthe media, as we demonstrate, then questions emerge about the extent to which such mediaare able to fulfill the normative role that they are typically ascribed.

The Role of the Media and the Importance of Ownership

Journalists are often trained to view their work’s purpose as providing citizens with theinformation needed to be “free and self-governing” and helping promote the conditionsrequired for accountability in a democratic system (Kovach & Rosenstiel, 2001, p. 12).Indeed, works in political communication suggest that the information provided by themedia can help improve the provision of public goods (Besley & Burgess, 2002) and

2 Allison M. Archer and Joshua Clinton

levels of accountability in government, as well as increase levels of political activityamong citizens (e.g., Prat & Strömberg, 2005; Prior, 2007; Snyder & Strömberg, 2010).In addition, a large literature in political science examines the power of media to primeconsiderations (e.g., Iyengar & Kinder, 1987) or frame an event in a manner that affectspublic opinion by influencing how citizens process information (e.g., Chong & Druckman,2007; Nelson, Clawson, & Oxley, 1997).

Such literatures explore the consequences of media coverage, but they do not explorethe processes by which the media choose which information to present to citizen con-sumers. The decision of which issues to cover is essential for understanding the media’srole in shaping citizens’ awareness, and such coverage decisions are arguably influencedby a myriad of factors—including both taste and profit motives.

Many explore important questions about how events are covered and whether mediaoutlets are “biased” in their coverage (e.g., Gentzkow & Shapiro, 2010; Groseclose & Milyo,2005; Ho & Quinn, 2008). Whether events are covered in a slanted direction because ofsupply-side or demand-side forces is an ongoing debate (see, for example, the summary of Prat& Strömberg, 2013), and both theoretical and empirical evidence support both factors(Gentzkow, Shapiro, & Stone, 2016; Puglisi & Snyder, 2015). On the supply side, Baron(2006) notes that some organizations may allow for journalists to inject bias into their reportingas an alternative to higher salaries, and some find evidence of owners’ preferences affectingthe coverage of their respective newspapers to their own benefit (Gilens & Hertzman, 2000;Puglisi, 2011). Numerous works also argue that there are demand-side pressures created bypreferences to consume news consistent with beliefs (e.g., Chan & Suen, 2008; Gentzkow &Shapiro, 2006; Mullainathan & Shleifer, 2005; Stroud, 2008).

We take a step back to explore a different, but equally important issue—the discretioninvolved in choosing whether or not to even cover political issues and how that coverage isaffected by who owns the media. The many excellent examinations of whether coverage is“biased” are in some sense secondary to understanding the conditions under which politicalevents are even covered. Moreover, focusing on the extensiveness of political coverage ratherthan the nature of political coverage has the added virtue of contributing to our knowledge ofmedia effects while avoiding the many difficult debates about how best to measure bias.

Anecdotal evidence regarding the importance of media owners in shaping their newsoutlets’ content is plentiful. Since buying the Washington Post, Jeff Bezos—also thefounder and CEO of Amazon.com—has urged the Post newsroom to focus on andexperiment with projects that have the potential for large-scaled digital growth (Meyer,2014). In the 2016 election cycle, the New York Observer—owned by Donald Trump’sson-in-law Jared Kushner—endorsed Trump in the Republican primary despite initialpledges that it would not cover him. In fact, some argued that the New York Observer“did not hesitate to advance [Trump’s] agenda” and pointed out that the Observer ran “alengthy takedown” of Trump’s one-time opponent, Marco Rubio (Mahler, 2016).

More systematic and quantitative work in political communication has also documentedthe power of newspaper owners in influencing the coverage of their papers (see, e.g.,Bagdikian, 1997). For instance, the ownership structure (e.g., public versus private) appearsto affect the depth of issue coverage (Dunaway, 2008), attention given to the “horse-race”(Dunaway&Lawrence, 2015), and the tone of political news stories (Dunaway, 2013) becauseprivately ownedmedia often have different goals—and fewer leaders at the helm to impede theachievement of those goals—compared to public media corporations. Others have argued thatmedia owners with a vested interest in the coverage of certain topics (e.g., the Citizens Unitedruling or 1996 Telecommunications Act) cover those topics in a personally beneficial manner(Bailard, 2016; Gilens &Hertzman, 2000). Additional work suggests that theNYTemphasized

Changing Owners, Changing Content 3

issues “owned” by Democrats more during presidential campaigns when there was aRepublican incumbent president—a shift that was attributed to a desire of the paper’s ownersto influence the election (Puglisi, 2011).

Newspaper owners have also been known to publicize their views so that reporters areable to tailor their coverage without direct intervention (Chomsky, 1999). Chomsky (2006)also argues that the owners of the NYT have sometimes directly intervened by choosingtopics for stories and guiding how stories are framed. An analysis of the correspondencebetween Turner Catledge, the top editor at the NYT for nearly two decades, and formerNYT owner and publisher Arthur Hays Sulzberger reveals that between 1956 and 1962,Catledge received 156 story suggestions from Sulzberger and at times, Catledge wouldforward memos from Sulzberger as his own (Chomsky, 2006).

We examine the question of whether owners affect media coverage by focusing onRupert Murdoch’s purchase of the WSJ—the top-selling daily newspaper in the UnitedStates—in August 2007. Although the transaction involved a purchaser with a portfolio ofright-leaning media companies (Jamieson & Cappella, 2008), Murdoch was seeking to buya popular newspaper largely known for its front-page coverage of economic events andnews; thus, it was not obvious a priori that changes would occur. The paper’s relativesuccess (it had one of the highest daily circulations at the time) and the close connectionthat exists between economic issues and the financial reputation of the company owningthe WSJ (The Dow Jones Company) arguably lessened the incentives to change thecoverage of the WSJ and risk alienating its existing audience.

Although there are reasons to think that coverage in the WSJ would remain stable afterits purchase from the Bancroft family—which had controlled the paper since 1902—thereare also reasons to think that a change may have resulted because of either preference orprofit motives. To be clear, our analysis is able to identify whether a shift in coverage occurswith the change in ownership, but we can only speculate about the possible reasons for ashift. One possibility is a difference in taste: Perhaps the Bancroft family had a preferencefor focusing on economic news and this preference—again, seemingly justified by the factthat the WSJ continued to be the top-selling daily newspaper in the years leading up to its2007 sale—led them to focus primarily on economic issues and events. Suggestive evidenceof the importance of the Bancrofts’ personal convictions is exemplified by the strongresistance that some family members had to selling the WSJ to Murdoch because of howhe might run the paper even though Murdoch was offering a 67% premium over its shareprice (Tofel, 2011). In contrast, perhaps Murdoch had a difference of taste and a desire toincrease political coverage even if it risked subscriptions to the WSJ.

Alternatively, Murdoch may have thought that increasing political coverage would resultin increased profits—a view consistent with his experiences at other news outlets (e.g., FoxNews). Although Hamilton (2004) notes somemedia outlets have boosted their amount of softnews in an effort to revive disappointing ratings in an increasingly fragmented media era,others suggest that political content can attract audiences (see Prat & Strömberg, 2013 foroverview). Even more, from a business perspective, the WSJ’s niche economic reportingwould possibly allow for higher profit margins vis-à-vis other daily papers.

We can only evaluate the profitability of decisions in retrospect, but the evidencesuggests that circulation of the WSJ increased after its sale. Figure 1 plots the annualcirculation levels (as of September) of the NYT and WSJ over nearly 30 years, and if wecompare the rate of change from 2007 to 2008, there is some evidence that the latter’scirculations actually increased despite the financial crises (the WSJ’s circulation increasedby 0.006% while those of the NYT decreased by 3.58%). The NYT did experience steepincreases in their circulations in the later years of Figure 1 (i.e., post-2010), but that could

4 Allison M. Archer and Joshua Clinton

be due to many factors: Democrats’ losses in the 2010 midterms (Archer, in press), aslowly rebounding economy, or the establishment of an online paywall in 2011 thatremoved free, unrestricted access to their content online.

Research Design and Data

To explore whether a change in owners affects the issues covered by the media, we collectdata on every front-page story as well as information related to the emphasis given to eachstory—whether it occurs above the fold, whether it contains an associated photo, and thesize of the headline font—for periods before and after the ownership change.2 To helpdisentangle editorial discretion from coverage decisions, we collect similar data oneditorials that attribute authorship to each paper’s Editorial Board. Because the contentof the front page is arguably more affected by real-world events than editorials, comparingthe relative change in coverage across papers and across these parts of the newspaper helpsilluminate the nature of the change in focus that may result from a change in ownershipand emphasis.

We collect information on every story, article, or unsigned opinion appearing on eitherthe front page or the editorial page of the NYT and WSJ for every weekday in 2007 and 2008(excluding holidays on which the papers did not print), and to help identify the potentialimpact of the 2008 presidential election on coverage, we also collected the same data for themonths of March, July, and November 2004. For robustness, we also collect information onthe coverage of USA Today (USAT) and the Washington Post (WaPo) for selected weekswithin each month. These two papers not only differ in their baseline proclivity to coverpolitical events, but they also have stable ownership structures over this time period.

Six coders were tasked with reading and coding the front-page stories and editorials.The coders were trained together and asked to extract the headline(s) for each of these

Figure 1. Weekday circulation levels of NYT and WSJ.

Changing Owners, Changing Content 5

articles or editorials. Once extracted, each piece was then coded for markers of prominenceand the content of the news coverage.3,4 That is, both front-page stories and editorials wereclassified as mentioning the following political topics: public policy, a federal agency, thepresident, Congress, the courts, candidates or campaigns, and interest groups.

Our primary analysis uses a difference-in-differences design to compare the coverageof the NYT and WSJ over time. This is an appropriate comparison because both papers arelocated in New York City, and they compete for subscribers both nationally and locally. Inaddition, they are two of the top three most prominent papers of record with printcirculations (Alliance for Audited Media, 2013). Traditionally, the non-editorial coverageof the WSJ was characterized as more focused on economic issues than political ones, butthe political positions taken by the Editorial Board in editorials were often political anddistinctive from those taken by the Editorial Board at the NYT. In terms of the “bias” foreach, studies routinely identify the NYT as a leftward-leaning paper and the WSJ as arightward-leaning paper (e.g., Gentzkow & Shapiro, 2010).

To identify the impact of the purchase of the WSJ by Rupert Murdoch, we use ournewly collected data to compare the extent to which political coverage differs alongseveral metrics: editorials, political stories on the front page, and political stories on thefront page that are emphasized in various ways. By examining the difference in coveragebetween the two papers over time, we are able to hold the larger political-social-economicenvironment fixed without having to identify and control for every newsworthy eventbecause the same events will exert similar pressures on both papers, especially since bothare headquartered in New York City. That is, analyzing the difference in coverage overtime is important because it allows us to examine how papers respond to identicalcircumstances while avoiding the impossible task of trying to control for all relevantevents that occurred during the time period of our analysis.

The second “difference” in the design involves the comparison of how the differencein coverage patterns for the two papers changed following the sale of the WSJ to RupertMurdoch in August 2007. To the extent that the difference in the coverage changesfollowing this sale, we can interpret the change as being due to the change in ownership.To be clear, both papers in the comparison may change their coverage in response to thepurchase of the WSJ—perhaps in response to the changing competitive landscape—but theeffect we identify is whether the WSJ experienced an even larger change.

Changing Owners, Changing Coverage?

Figure 2 plots the proportion of political stories and editorials appearing in the weekday papersof theWSJ and NYT across time—the top row graphs the trends for front-page political storiesand editorials by week while the bottom row depicts these trends by month.

It is immediately clear that the difference between the proportion of political stories onthe front pages of the WSJ and the NYT decreases dramatically after Murdoch’s purchaseof the WSJ. Prior to August 2007 (excluding our 2004 data), only 29.86% of the WSJ’sfront-page coverage focused on politics on average, compared to an average of 59.83% onthe NYT’s front pages. After August 2007, the mean percentage of the WSJ’s front-pagestories jumped to 47.14% compared to 66.59% for the NYT.

While some of this may be due to events involving the 2008 presidential election, theelection alone cannot explain the large variation we identify. By way of comparison, forthe selected months we coded in the 2004 presidential election year, the average percen-tage of front-page political stories in the WSJ was only 38.52% compared to 59.95% forthe NYT. The fact that the WSJ’s political coverage in 2004 was only slightly higher than it

6 Allison M. Archer and Joshua Clinton

was in 2007 prior to Murdoch’s purchase suggests that the 2008 presidential electioncannot fully account for the dramatic shift.

The right column of Figure 2 repeats the same analysis for editorials and reveals a starklydifferent pattern. The editorials for both papers focused nearly exclusively on political issues priortoMurdoch’s purchase of theWSJ—before August 2007, 78.86% and 81.89% of editorials werepolitical in the WSJ and NYT, respectively—and the relative number of politically themededitorials in the WSJ did not substantially increase after its purchase. (After Murdoch tookover, 86.83% and 87.62% of theWSJ and NYT’s editorials were political.)5

Simple t-tests examining the differences in the proportions of political content on thefront and editorial pages of the two papers over time suggest the political content on theeditorial pages did not meaningfully change in response to the change in ownership;however, the gap in political content on the front page shrank significantly in bothsubstantive and statistical terms (from –.30 to –.19; p = .001).6

0.0

0.2

0.4

0.6

0.8

1.0

Front Page Political Stories (Weekdays) by Week

Week

Pro

port

ion

2004 2007 2008 0.0

0.2

0.4

0.6

0.8

1.0

Political Op Eds (Weekdays) By Week

Week

Pro

port

ion

NYTWSJ

2004 2007 2008

0.0

0.2

0.4

0.6

0.8

1.0

Front Page Political Stories (Weekdays) By Month

Month

Pro

port

ion

Mar Ju

lNov Ja

nFe

bM

ar AprM

ay Jun Ju

lAug Sep Oct

Nov Dec Jan

Feb

Mar Apr

May Ju

n Jul

Aug Sep OctNov Dec

2004 2007 2008 0.0

0.2

0.4

0.6

0.8

1.0

Political Op Eds (Weekdays) By Month

Month

Pro

port

ion

Mar Ju

lNov Ja

nFe

bM

ar AprM

ay Jun Ju

lAug Sep Oct

Nov Dec Jan

Feb

Mar Apr

May Ju

n Jul

Aug Sep OctNov Dec

2004 2007 2008

Figure 2. Proportion of political front-page stories and editorials across time.Note: Figures include smoothing curve line fitted by LOESS with smoothing parameter set to 0.4and 0.5 for weekly and monthly data, respectively.

Changing Owners, Changing Content 7

To characterize the change more completely and better control for possible unob-served factors, we employ regression models to model the differences in coverage betweenthe NYT and WSJ, using an indicator for the time after the sale of the WSJ—Murdoch—andtime fixed effects to control for variation in political events and remove temporally basedconfounders. Specifically, we estimate regressions of the form:

WSJt � NYTtð Þ ¼ αþ β �Murdochþ θt (1)

where WSJt and NYTt are the proportion of front-page stories or editorials in time period tdealing with politics in each paper, α is the average difference between the WSJ’s andNYT’s political coverage, β is the change in that difference following the first full weekafter Murdoch’s purchase of the WSJ (which began on August 5, 2007), and θt are fixedeffects at either the calendar week (t = 1…51) or calendar month (t = 1…11) level.7 Notethat positive values of β represent an increase in the political coverage of the WSJ relativeto the NYT post-Murdoch and negative values represent the opposite.

For clarity, Table 1 reports the results using 2007 and 2008, but the findings are robustto the inclusion of data from 2004 in the time period before Murdoch’s tenure.8

Confirming the patterns evident in Figure 2, the first column of Table 1 reveals thatthe relative percentage of political editorials did not significantly change followingMurdoch’s purchase of the WSJ—likely because both papers were already publishingsuch a high proportion of political editorials.

In contrast, the results reported in columns 2 through 4 reveal meaningful and sizablechanges in the WSJ’s front-page political coverage regardless of whether calendar week orcalendar month fixed effects are used to account for news cycle effects. While the WSJcontained nearly half as many front-page political stories as the NYT prior to the purchaseby Murdoch, that difference fell considerably post-purchase. Prior to August 2007, therewas a 30-percentage-point average gap between the NYT and the WSJ, but that gap shrunkby 10 percentage points on average (roughly 35%) after Murdoch’s purchase. While it istrue that each paper boosted its political coverage in 2008 relative to 2007 because of thepresidential election, the WSJ did so at a rate (57.87%) far greater than the NYT (11.30%).

Table 1Differences in political content

Editorials(by Week)

Front-Page(by Week)

Front-Page(by Week)

Front-Page(by Month)

(1) (2) (3) (4)

Murdoch .02 .10*** .15*** .15**(Std. Err.) (.03) (.03) (.04) (.06)Intercept –.03 –.30*** –.62*** –.47***

(.02) (.03) (.12) (.10)Weekly fixed effects No No YesMonthly fixed effects YesN 104 104 104 24R2 .008 .09 .50 .47

**p < .05. ***p < .01, two-tailed.

8 Allison M. Archer and Joshua Clinton

To measure whether the attention given to political events also changed in response tothe change in ownership, Figure 3 plots the proportion of political stories (by week) abovethe fold of the front page for the NYT and WSJ. The patterns suggest that in the selectmonths for 2004 and most of 2007, the NYT had a higher proportion of front-page politicalstories above the fold than the WSJ, but that this difference decreased substantially in2008. Indeed, 44.84% of front-page stories in the NYT, on average, were both political andabove the fold in the select months for 2004, compared to only 27.60% for the WSJ.Between January 2007 and August 2007, the percentage of political, above-the-fold front-page stories was 41.31% in the NYT and 14.61% in the WSJ—percentages that changed to37.19% for the NYT and 27.49% for the WSJ following Murdoch’s purchase of the WSJ.

Table 2 confirms the significance of this change by predicting the difference in politicalfront-page stories located above the fold (operationalized asWSJminusNYT) before versus afterMurdoch. The results reveal a sizable increase in the emphasis of political stories following thesale of theWSJ; the NYT went from having 26.7 percentage points more political stories abovethe fold than theWSJ (from January toAugust 2007) to just 9.69 percentage points more than theWSJ afterMurdoch’s purchase.While it is hard to determine definitively whether this is due to anincrease in the number of political stories, the emphasis given to political stories, or both, it is clearthat relatively more political stories were being promoted after Murdoch took over theWSJ.9

Columns 2 and 3 of Table 2 explore the results for other measures of emphasis: Thepercentages of stories that were both political and in the top-right position of the twopapers’ front pages (placement in the top right of the front page has historically signaledthat the article is the most important of the day), and the percentage of front-page politicalstories with a photo. While the gap between the differences in these percentages closesover time, such changes are neither substantively nor statistically impressive. These effectsare not completely attributable to more political events occurring post-purchase thanbefore because the results of all three models are robust to the inclusion of data fromthe 2004 presidential election year (see Table A3 of supplemental Appendix A).

Figure 3. Proportion of political stories above the fold.Note: Figure includes smoothing curve line fitted by LOESS with smoothing parameter set to 0.4.

Changing Owners, Changing Content 9

The results of Tables 1 and 2 demonstrate that political coverage increased in responseto a change in ownership at the WSJ, but shed no light onto what was being included inthat additional coverage. To do so, we compare the changes in the frequencies of keywordsrelated to politics within the headlines of editorials and front-page stories over time.Table 3 presents the percentage increase in the frequency of each word for both papers’editorial (columns 1 and 2) and front-page (columns 3 and 4) headlines from beforeMurdoch’s purchase of the WSJ to after.10

Similar to Wagner and Collins (2014), Table 3 reveals clear evidence of the WSJ’seditorial page headlines increasing their mentions of the word “Democrat(s)” and doing soat a rate much greater than NYT editorials (650% versus 100%). In addition, theWSJ editorialpage headlines increased mentions of Obama at a rate much larger than any of the other words

Table 2Differences in emphasis of political front-page stories by week

Stories Above the Fold Stories in Top Right Stories With a Photo

(1) (2) (3)

Murdoch .16*** .03 .04(Std. Err.) (.03) (.02) (.03)Intercept –.49*** –.12** –.15*

(.10) (.05) (.08)Weekly fixed effects Yes Yes YesN 104 104 104R2 .66 .49 .58

*p < .10. **p < .05. ***p < .01, two-tailed.

Table 3Percentage change in frequency of keywords in headlines before and after Murdoch

% Change inNYT Editorial

% Change inWSJ Editorial

% Change in NYTFront-Page

% Change in WSJFront-Page

(1) (2) (3) (4)

Democrat 100 650 69.70 155.56Republican 300a 0 100 200GOP 100a 100 44.44 200Bush 50 500 41.67 42.86Pelosi 100a 700 –100 0Obama 900a 2,650 1,038.46 3,150McCain 900a 850 825 3,100Clinton 200a 133.33 435.71 1,266.67

aDenotes values that equaled 0 before Murdoch’s purchase of the WSJ but were transformed toequal 1 so that percentage changes could be calculated. Cell entries are the percentage change in thefrequencies of each word in the headlines of both editorials (columns 1 and 2) and front-page stories(columns 3 and 4) from before to after Murdoch’s purchase. Frequency counts include both thesingular and plural (or possessive) forms of each word.

10 Allison M. Archer and Joshua Clinton

listed in Table 3. More strikingly, there was a larger percentage increase for all of the keypolitical words in Table 3 on theWSJ’s front-page headlines compared to theNYT’s front-pageheadlines. The largest relative increases for the WSJ occurred in its front-page mentions ofpresidential candidates running in 2008, particularly Barack Obama and John McCain. Whilethis analysis cannot tell us the tone of the coverage of each party, candidate, or key figure, itdoes provide insight into one of the ways in which theWSJ boosted its front-page and editorialpolitical content: by focusing more on the candidates for office, particularly Obama.

Is It All About the Campaign?

One potential concern with the results thus far is that the increase in the WSJ’s front-pagepolitical content between 2007 and 2008 may have been due to the presidential election.The fact that similar results emerge when we include coverage of the 2004 presidentialelection provides some evidence that this is not likely to be the case, but we further probethe robustness of our findings by repeating the analysis after removing all stories that focuson the campaign and candidates. That is, this section asks the following: Is there anincrease in stories focused on public policy, a federal agency, the president, Congress, thecourts, or interest groups but that do not mention presidential candidates or campaigns?

Table 4 reports the results. Similar to before, the results in columns 1 and 2 suggestthat there are not significant changes in the difference in political editorials between thetwo papers over time. The results of column 3, however, reveal that there are significantchanges in the difference between the two papers’ political front-page content before andafter Murdoch’s acquisition of the WSJ. Column 3 of Table 4 reports a positive andsignificant (b = .11; p = .001) bivariate relationship between the difference in the twopapers’ political front-page stories and the sale of the WSJ (Murdoch). Including calendarweek fixed effects to control for time-varying content in column 4 only increases themagnitude of the estimated effect (b = .16; p = .00).

Moreover, consistent with the impact being due to the change of ownership at theWSJ, the decreased difference is due almost entirely to changes at the WSJ: the WSJshifted from 29.15% of front-page stories having political content (excluding mentions of

Table 4Differences in political content excluding stories about campaign

Editorials(by Week)

Editorials(by Week)

Front-Page(by Week)

Front-Page(by Week)

(1) (2) (3) (4)

Murdoch .03 .02 .11*** .16***(Std. Err.) .03 .03 .03 .04Intercept –.03 –.02 –.29*** –.61***

.02 .09 .03 .12Weekly fixed

effectsNo Yes No Yes

N 104 104 104 104R2 .009 .52 .10 .50

Note. Stories about campaigns or candidates excluded from coding of political stories.***p < .01, two-tailed.

Changing Owners, Changing Content 11

campaigns and candidates) before Murdoch’s purchase to 45.58% after his purchase.During the same time periods, the NYT experienced a much smaller increase from58.28% political (non-campaign) front-page stories to 63.16%.

Comparing Changes in the WSJ to Changes in Other Papers

To ensure the robustness of our characterization, we also compare the coverage of the WSJto two other prominent papers with a national audience, USA Today (USAT) and theWashington Post (WaPo). The Graham family famously owned the WaPo for 80 years upuntil Jeff Bezos’s purchase of the paper in fall 2013 (Farhi, 2013), and USAT has beenowned by the Gannett Company since its initial development and publication in 1981 and1982, respectively.11

We collect information on the number of political stories and total stories on the frontpages of these additional papers for the first full week of every month we examine (i.e., forall of 2007 and 2008 and the months of March, July, and November 2004), and we codethe text of each front-page story. (Supplemental Appendix D graphs the time series ofcoverage for the four papers.)

Unlike the NYT, the content of USAT is often less politically focused, which means thatthere is more room for an increase in coverage because of the 2008 election. We initiallyreplicate our previous analyses by differencing the average weekly amount of political cover-age in the WSJ relative to the WaPo and USAT; while the percentage of the WSJ’s politicalfront-page stories in the first week of every month increased by 21.27 percentage points, onaverage, after Murdoch’s purchase of the paper, the USAT and WaPo only experiencedincreases of 3.45 and 1.93 percentage points, respectively, during the same time period.Columns 1 and 3 of Table 5 shed light on the significance of those differences in changesamong the WSJ and each of the two new comparison papers. In each case, the bivariateregressions suggest the gaps between the WSJ and the other papers decreased at a significantrate. Columns 2 and 4 of Table 5 confirm that the 18- and 19-point decreases in theWSJ’s gapin political coverage relative to theUSAT andWaPo, respectively, remain substantively similar

Table 5Differences in political content: Comparisons with USA Today and Washington Post

Front-Pagevs. USAT

Front-Pagevs. USAT

Front-Pagevs. WaPo

Front-Pagevs. WaPo

(1) (2) (3) (4)

Murdoch .18*** .16* .19*** .15(Std. Err.) (.06) (.08) (.06) (.10)Intercept –.55*** –.54*** –.61*** –.53***

(.05) (.13) (.05) (.16)Fixed

effectsNo Yes No Yes

N 24 24 24 24R2 .27 .57 .30 .43

Note. Data consist of coding for first full week in each month. Results are substantively similarincluding data from selected months in 2004.*p < .10. ***p < .01, two-tailed.

12 Allison M. Archer and Joshua Clinton

after using fixed effects for calendar months to control for changing news cycles over time.The p-value for the key coefficient in column 4 comparing theWSJ toWaPo is onlymarginallysignificant (p = 0.15, two-tailed) but it is in the anticipated direction and it reaches conven-tional levels of significance in a one-tailed test.

We can also pool the results and use a model that explores the impact of the change inownership on coverage by estimating a specification predicting the percentage of politicalstories as a function of idiosyncratic paper effects and time using variations on thefollowing specification:

yit¼ αþ β1IWSJ þ β2IUSAT þ β3IWAPO þ β4 IWSJ�Murdochð Þþ β5 IUSAT�Murdochð Þþ β6 IWAPO�Murdochð Þþ τt

where IWSJ, IUSAT, and IWAPO are indicators for whether paper i is the WSJ, USAT, or WaPo,respectively (the excluded paper is the NYT), Murdoch is an indicator if the date t is after thesale of theWSJ, and τt captures the effects of time modeled using either calendar month fixedeffects or a time trend. Critical for our investigation are the interaction effects β5, β6, and β7,which capture how coverage changes following the sale of the WSJ. Although we includepaper and time fixed effects to account for systematic changes, we also cluster the standarderrors by paper to allow for different error structures using the wild cluster bootstrap(Cameron, Gelbach, & Miller, 2008) when feasible (note that we did not need to do so inprior analyses because the dependent variable of interest was the difference in coverage).12

The results of this pooled analysis are presented in Table 6. Specifications (1) and (2) inTable 6 compare the coverage of theWSJ to a pooled comparison of theNYT, USAT, andWaPowith (specification 2) and without (specification 1) calendar month fixed effects to control fordifferences in the political news cycle across a given year. The results, unsurprisingly, confirmearlier findings: while theWSJ is systematically less likely to cover politics by 44% on averagerelative to the other papers, the percentage of front-page political coverage post-purchaseincreases between 13% (specification 2) and 17% (specification 1). Moreover, while a smallfraction of this increase affects all papers, 12% of each estimated increase is unique to theWSJ.Digging deeper and estimating a model with paper fixed effects and separate effects by paperreveals continued support for the claim that the WSJ experienced a substantively and statis-tically distinctive change in coverage after its sale to Rupert Murdoch. Prior to the sale, theWSJ front page had 28% fewer political stories than theNYT, but the difference shrank by 16%following the sale. While theWSJ is still less likely to cover political events than the NYT andother major papers—perhaps owing to its legacy as the national economic paper of record—there is, nonetheless, clear evidence that the content shifted dramatically following its sale toRupert Murdoch. Moreover, the increase in political coverage was not largely attributable totrends that produced a shift in the other papers, as most of the increase we observe is unique tothe WSJ.13

Taken together, examining the shifts in the political content on the front page of theWSJ against those on the front pages of the NYT, USAT, and the WaPo provides consistentevidence that the WSJ experienced increases in its political content at a rate that faroutpaced its peer group of nationally circulated newspapers. Moreover, this increase isgreater than what might be expected from the 2008 election alone because it is greater thanthe changes we observe in 2004, and evidence of the shift remains even after excludingstories that focus on the 2008 campaign. As a result, our findings suggest that owners canmeaningfully influence their outlets’ coverage—particularly the news that makes the frontpages of their papers—and that better understanding the incentives and impact of owners’decisions is important for understanding the role of the media.

Changing Owners, Changing Content 13

Conclusion

Democracies rely on the “fourth estate” to provide critical coverage of the elitesrepresenting the masses. It is well-known that the media have powerful effects throughframing, priming, and setting the agenda for political and social discourse, but char-acterizations of the processes leading to the selection of some content over other iscomparatively rare. We collect and analyze new data from real-world newspaper cover-age to demonstrate that a powerful way that agenda-setting can occur is through theinfluences of media owners.

We find evidence consistent with the claim that who owns the news matters when itcomes to the content on the front pages of the WSJ, one of the top newspapers in America.We show that the amount of front-page political coverage in the WSJ—and also theemphasis that is given to political events—increased markedly after the sale of the paperto Rupert Murdoch relative to the change in front-page political coverage in the NYT,USAT, and the WaPo. While our analysis does not probe the important question of howthose political events were covered, the fact that more politics was being covered issuggestive of the impact of ownership on the stories being covered by the press.

Table 6Pooled analysis of political coverage

Front Page (by Week) Front Page (by Week) Front Page (by Week)

(1) (2) (3)

WSJ –.44 –.44 –.28[–.55, –.33] [–.55, –.33] [–.37, –.19]

WaPo .28[.20,.38]

USAT .21[.12,.30]

Murdoch X .12 .12 .10WSJ [.10, .14] [.10, .14] [–.02, .21]Murdoch X –.05WaPo [–.17, .06]Murdoch X −.02USAT [–.13, .10]Murdoch .05 .01 .07

[.03, .07] [–.02, .04] [–.01, .16]Intercept .78 .81 .62

[.67, .89] [.78, .85] [.55, .68]Month fixed No Yes NoeffectsN 108 108 108R2 .59 .61 .78

Notes. Ninety-five percent confidence intervals reported. Specifications (1) and (2) use wildbootstrap standard errors (Cameron et al., 2008) clustered by paper. Data consist of the first fullweek of each month in 2007 and 2008 as well as three months from 2004. Murdoch is an indicatorfor weeks following the sale of the WSJ in August 2007.

14 Allison M. Archer and Joshua Clinton

Our findings highlight the essential tension facing journalists and the “free press.”Whilethe media are free from governmental interference, they are still a political institution in thatthey are entrenched in policymaking and elections; that is, they are an institution “that actsapart from, but in the middle of, politics” (Esser et al., 2001, p. 22). Even more, they are notfree frommarket forces and the necessity of delivering profits or other benefits to their owners.And although this particular case may have led to increased political coverage in a majornational newspaper—a result that could be beneficial at least at face value—not all instances ofownership influence lead to the same result. For instance, conservative billionaire and politicalinfluencer Sheldon Adelson’s purchase of the Las Vegas Review-Journal in 2015 led tochanges in the paper’s journalistic practices that undercut editorial freedom (particularlyfreedom in covering Adelson’s business, legal, and political affairs) to the point wherereporters and editors left the paper (Peralta, 2015; Wagner, 2016). Furthermore, Murdoch’sown news outlets in the United Kingdom are known to have pointed political agendas and tohave boasted about their influence over elections (Douglas, 2010). Future workmight examinethe structure of news outlets’ ownership in cross-national contexts to deepen our under-standing of these key findings as well as how differences in political institutions and themedia environment affect one another.

Given the critical role that the press occupies in creating informed citizens and holdingelected officials responsible for their actions, the impact of changing ownership we documentsuggests many critical questions remain regarding the incentives for why the press chooses tocover the stories and events that it chooses to cover. Future qualitative and quantitative workmight delve deeper into the personal motivations of owners’ decisions to affect news coverage(e.g., in support of one policy or not as in Bailard, 2016; Gilens & Hertzman, 2000) becausethis will help us better understand the types of incentives that lead to particular changes innews coverage, as well as the specific effects such coverage changes have on public opinionabout various policies, practices, and politicians.

Supplemental Material

Supplemental data for this article can be access on the publisher’s website at https://doi.org/10.1080/10584609.2017.1375581

Notes

1 Moreover, we are also able to use our daily content data to estimate the amount of time that ittook for those changes to reveal themselves. The data suggest that the change in politicalcoverage by the WSJ was not fully evident until December 2007—four months after Murdochtook over.

2 Prior studies have shown that the front page of the WSJ covered business less in 2011 relativeto 2007 both in absolute terms and also relative to the NYT (Pew Research Center, 2011), butsuch analyses focus almost entirely on post-purchase coverage and face difficulties in control-ling for potential changes in the news cycle and the media environment.

3 During training, coders were given sample articles to code. Inter-coder reliability was rather highfor items, including above the fold (kappa = 0.85), top-right front-page position (kappa = 0.96),mentioning the president (kappa = 0.73), and mentioning a candidate or campaign (kappa = 0.57).Reviewing the differential codings for campaign mentions reveals that when the politics of acurrent event were being discussed and the positions of candidates were also mentioned as an asidethere were differences of opinion as to whether the emphasis of the story was on current policy-making or presidential campaign. For robustness, we estimated models with and without thismeasure and the results are reassuringly substantively identical.

Changing Owners, Changing Content 15

4 For front-page news stories, the following were coded as markers of emphasis: whether or notthe headline was in bold font, the size of the headline font (largest on page), whether or not theheadline is in all caps, if the story is above the fold, if the story is in the top-right position, andif a photo accompanied the story.

5 That said, while we focus on the amount of political coverage, others have found that editorialsin the WSJ became less supportive of economic government intervention and more likely tomention Democrats post-transition. The WSJ’s editorial page also became more negativetoward Democrats and more positive toward Republicans. Similar changes did not occur inthe editorial pages of the NYT or Washington Times, papers with liberal and conservativeeditorial views, respectively, who did not experience changes in ownership (Wagner & Collins,2014). Kahn and Kenney (2002) suggest there is a relationship between the explicit politicalpositioning of newspaper editorials and the papers’ supposedly objective news content. Otheranalyses also find a positive relationship between bias scores for newspaper editorials and threemeasures of news reporting (Ho & Quinn, 2008).

6 See supplemental Appendix A for results.7 Because of the misalignment of full weeks in 2007 and 2008, December 29–31, 2008, are

dropped from analyses so that only full weeks are included.8 See supplemental Appendix A for analyses including 2004 data.9 This effect is robust to controlling for the difference in the number of political front-page

stories for each paper, which does suggest that not only was the WSJ more likely to coverpolitics on its front page, but it was also more likely to emphasize that political content.Analyses can be made available upon request.

10 See supplemental Appendix B for the raw frequencies of words mentioned before and afterMurdoch’s purchase of the WSJ.

11 There was a change in WaPo publisher in February 2008 and in executive editor in July 2008(see Ahrens, 2008).

12 This data set has an N = 108: 27 weeks (one for each of the three months used previously from2004, as well as one for each month in 2007 and 2008) for each paper. Our ability to estimateclustered standard errors is limited for some specifications given the small number of cases perpaper and the inclusion of interactions alongside paper and time fixed effects.

13 While the 95% confidence interval for the shift that is unique to the WSJ ranges from ––0.02 to0.21, with a point estimate of 0.10 recall this assumes that the transition occurs immediatelyupon the sale of the WSJ. If we allow a few weeks for an effect to emerge, the point estimate ofthe interaction increases to 0.20 with a 95% confidence interval ranging from .10 to .32.

Acknowledgements

We thank Andrew Mah and Jasmine Shin for their outstanding research assistance. We alsothank Sam Adkisson, Peter Liu, and Caylyn Perry for serving as excellent research assistantson this project. Archer acknowledges support from the National Science FoundationGraduate Research Fellowship Program under Grant Numbers 0909667 and 1445197.

ORCID

Allison M. Archer http://orcid.org/0000-0002-8767-8898Joshua Clinton http://orcid.org/0000-0003-1683-9479

References

Ahrens, F. (2008). Post Co. names Weymouth media chief and publisher. Retrieved from http://www.washingtonpost.com/wp-dyn/content/article/2008/02/07/AR2008020701162.html

16 Allison M. Archer and Joshua Clinton

Alliance for Audited Media. (2013). Top 25 newspapers for March 2013. Retrieved from http://auditedmedia.com/news/blog/top-25-us-newspapers-for-march-2013.aspx

Archer, A. M. N. (in press). Political advantage, disadvantage, and the demand for partisan news.The Journal of Politics.

Bagdikian, B. H. (1997). The media monopoly. Boston, MA: Beacon Press.Bailard, C. S. (2016). Corporate ownership and news bias revisited: Newspaper coverage of the

Supreme Court’s Citizen United ruling. Political Communication, 33(4), 583–604.Baron, D. P. (2006). Persistent media bias. Journal of Public Economics, 90(1), 1–36. doi:10.1016/j.

jpubeco.2004.10.006Besley, T., & Burgess, R. S. L. (2002). The political economy of government responsiveness: Theory

and evidence from India. Quarterly Journal of Economics, 117(4), 1415–1452. doi:10.1162/003355302320935061

Cameron, A. C., Gelbach, J. B., &Miller, D. L. (2008). Bootstrap-based improvements for inference withclustered errors. The Review of Economics and Statistics, 90(3), 214–247. doi:10.1162/rest.90.3.414

Chan, J., & Suen, W. (2008). A spatial theory of news consumption and electoral competition.Review of Economic Studies, 75(3), 699–728. doi:10.1111/j.1467-937X.2008.00495.x

Chomsky, D. (1999). The mechanisms of management control at the New York Times. Media,Culture & Society, 21(5), 579–599. doi:10.1177/016344399021005001

Chomsky, D. (2006). “An interested reader”: Measuring ownership control at the New York Times.Critical Studies in Media Communication, 23(1), 1–18. doi:10.1080/07393180600570659

Chong, D., & Druckman, J. N. (2007). Framing theory. Annual Review of Political Science, 10, 103–126. doi:10.1146/annurev.polisci.10.072805.103054

Demsetz, H., & Lehn, K. (1985). The structure of corporate ownership: Causes and consequences.Journal of Political Economy, 93(6), 1155–1177. doi:10.1086/261354

Douglas, T. (2010). Analysis: Murdoch and media ownership in UK. Retrieved from http://www.bbc.com/news/uk-12062176

Dunaway, J. (2008). Markets, ownership, and the quality of campaign news coverage. The Journal ofPolitics, 70(4), 1193–1202. doi:10.1017/S0022381608081140

Dunaway, J. (2013). Media ownership and story tone in campaign news. American PoliticsResearch, 41(1), 24–53. doi:10.1177/1532673X12454564

Dunaway, J.,&Lawrence, R.G. (2015).What predicts the game frame?Media ownership, electoral context,and campaign news. Political Communication, 32(1), 43–60. doi:10.1080/10584609.2014.880975

Esser, F., Reinemann, C., & Fan, D. (2001). Spin doctors in the United States, Great Britain, andGermany: Metacommunication about media manipulation. The International Journal of Press/Politics, 6(1), 16–45.

Farhi, P. (2013).Washington Post closes sale toAmazon founder Jeff Bezos. Retrieved from https://www.washingtonpost.com/business/economy/washington-post-closes-sale-to-amazon-founder-jeff-bezos/2013/10/01/fca3b16a-2acf-11e3-97a3-ff2758228523_story.html?utm_term=.dd628236956b

Gentzkow, M., & Shapiro, J. M. (2006). Media bias and reputation. Journal of Political Economy,114(2), 280–316. doi:10.1086/499414

Gentzkow, M., & Shapiro, J. M. (2010). What drives media slant? Evidence from U.S. dailynewspapers. Econometrica, 78(1), 35–71. doi:10.3982/ECTA7195

Gentzkow, M., Shapiro, J. M., & Stone, D. (2016). Media bias in the marketplace: Theory. In S. P.Anderson, J. Waldfogel, & D. Strömberg (Eds.), Handbook of media economics (vol. 2, pp.623–645). North-Holland, Amsterdam, Netherlands.

Gilens, M., & Hertzman, C. (2000). Corporate ownership and news bias: Newspaper coverage of the 1996Telecommunications Act. The Journal of Politics, 62(2), 369–386. doi:10.1111/0022-3816.00017

Groseclose, T., & Milyo, J. (2005). A measure of media bias. The Quarterly Journal of Economics,120(4), 1191–1237. doi:10.1162/003355305775097542

Hamilton, J. (2004). All the news that’s fit to sell: How the market transforms information into news.Princeton, NJ: Princeton University Press.

Ho, D. E., & Quinn, K. M. (2008). Measuring explicit political positions of media. QuarterlyJournal of Political Science, 3(4), 353–377. doi:10.1561/100.00008048

Changing Owners, Changing Content 17

Iyengar, S., & Kinder, D. R. (1987). News that matters: Television and American opinion. Chicago,IL: University of Chicago Press.

Jamieson, K. H., & Cappella, J. N. (2008). Echo chamber: Rush Limbaugh and the conservativemedia establishment. Oxford, UK: Oxford University Press.

Kahn, K. F., & Kenney, P. J. (2002). The slant of the news: How editorial endorsements influencecampaign coverage and citizens’ views of candidates. American Political Science Review, 96(2),381–394. doi:10.1017/S0003055402000230

Kovach, B., & Rosenstiel, T. (2001). The elements of journalism: What newspeople should know andthe public should expect. New York, NY: Three Rivers Press.

Larcinese, V., Snyder, J., & Puglisi, R. (2011). Partisan bias in economic news: Evidence on theagenda setting behavior of US newspapers. Journal of Public Economics, 95(9–10), 1178–1189. doi:10.1016/j.jpubeco.2011.04.006

Mahler, J. (2016). New York Observer endorses Donald Trump. Retrieved from http://www.nytimes.com/politics/first-draft/2016/04/12/new-york-observer-endorses-donald-trump/

Meyer, M. (2014). Brick by brick. Retrieved from http://www.cjr.org/cover_story/washington_post_-jeff_bezos.php

Mullainathan, S., & Shleifer, A. (2005). The market for news. The American Economic Review, 95(4), 1031–1053. doi:10.1257/0002828054825619

Nelson, T. E., Bryner, S. M., & Carnahan, D. M. (2011). Media and politics. In J. N. Druckman, D.P. Green, J. H. Kuklinski, & A. Lupia (Eds.), Cambridge handbook of experimental socialscience (pp. 201–213). Cambridge, UK: Cambridge University Press.

Nelson, T. E., Clawson, R. A., & Oxley, Z. M. (1997). Media framing of a civil liberties conflict and itseffect on tolerance. American Political Science Review, 91(3), 567–583. doi:10.2307/2952075

Peralta, E. (2015). Ending mystery, Adelson family says it bought the Las Vegas Review-Journal.Retrieved from http://www.npr.org/sections/thetwo-way/2015/12/17/460132950/ending-mys-tery-adelson-family-says-it-bought-the-las-vegas-review-journal

Pew Research Center. (2011). The Wall Street Journal under Rupert Murdoch. Retrieved from http://www.journalism.org/2011/07/20/wall-street-journal-under-rupert-murdoch/

Prat, A., & Strömberg, D. (2005). Commercial television and voter information. CEPR DiscussionPaper 4989.

Prat, A., & Strömberg, D. (2013). The political economy of mass media. In D. Acemoglu, M.Arellano, & E. Dekel (Eds.), Advances in economics and econometrics: Theory and applica-tions, proceedings of the tenth world congress of the econometric society. Cambridge, UK:Cambridge University Press.

Prior, M. (2007). Post broadcast democracy: How media choice increases inequality in politicalinvolvement and polarizes elections. Cambridge, UK: Cambridge University Press.

Puglisi, R. (2011). Being the New York Times: The political behavior of a newspaper. B.E. Journal ofEconomic Analysis & Policy, 11(1). doi:10.2202/1935-1682.2025

Puglisi, R., & Snyder, J. M. (2015). Empirical studies of media bias. In S. P. Anderson, J. Waldfogel,& D. Strömberg (Eds.), Handbook of media economics (pp. 647–667). North-Holland,Amsterdam, Netherlands.

Snyder, J. M., & Strömberg, D. (2010). Press coverage and political accountability. Journal ofPolitical Economy, 118(2), 355–408. doi:10.1086/652903

Stroud, N. J. (2008). Media use and political predispositions: Revisiting the concept of selectiveexposure. Political Behavior, 30(3), 341–366. doi:10.1007/s11109-007-9050-9

Tofel, R. (2011). FormerWall St. Journal owners: “We wouldn’t have sold if we had known.” Retrievedfrom https://www.theguardian.com/media/2011/jul/13/wall-st-journal-murdoch-bancroft

Wagner, L. (2016). More journalists leaving Las Vegas Review-Journal after sale to billionaire.Retrieved from http://www.npr.org/sections/thetwo-way/2016/05/09/477423367/more-journal-ists-leaving-las-vegas-review-journal-after-sale-to-billionaire

Wagner, M. W., & Collins, T. P. (2014). Does ownership matter? The case of Rupert Murdoch’spurchase of the Wall Street Journal. Journalism Practice, 8(6), 758–771. doi:10.1080/17512786.2014.882063

18 Allison M. Archer and Joshua Clinton